Privacy Preserving Data Publishing of Multiple Sensitive Attributes by using Various Anonymization Techniques

Abstract

The enhancement in digital era speeds up the process of aggregating the massive amount of information from various sectors of governments, diverse sections of healthcare unit, multiple organizations as well as from individuals. This aggregated data’s release is essential for the betterment of researchers, varied occupations, and individuals etc. This gives rise for necessitate releasing and exchanging of assembled data. However, when information is in native form, it carries some crucial sensitive facts about human beings and/or organizations. If such information is disclosed, personal and/or organizational privacy may be threatened. Therefore, Privacy Preserving Data Publishing (PPDP) comes up with tools and techniques which describe how to publish valuable facts along with its privacy protection. Thus, it is inevitable to alter the data before its release with the aim to persist its privacy without jeopardize its utility. This is achieved by varied anonymization schemes. In point of fact, datasets comprise of distinct kinds of Multiple Sensitive Attributes (MSAs) (which can be numerical and/or categorical). Anonymization done for only Single Sensitive Attribute is not having any importance in practical scenarios. On that account, it is significant that, while operating the highly dimensioned data, the association amidst these MSAs is sustained along with the efficient privacy preservation of Mixed (numerical as well as categorical) MSAs. This paper concentrates mainly on analysing different schemes proposed in literature for PPDP of MSAs.